Jan Karremans, Director of Sales Engineering at EnterpriseDB
Disruption has come to the role of the database administrator (DBA). Automation is eating up tasks typically completed by a DBA – and the technology is evolving to become ever smarter, ever more capable.
This automation has sparked concern that humans will inevitably become redundant in running databases. Already, cloud services have threatened to render the role obsolete. Now, with automation infiltrating more and more areas of database administration, many have predicted that DBAs will not survive the storm of change.
Though it might seem cloud computing and agile development could wipe out the traditional DBA, in fact, the role is changing. Rather than lurking in the data centre doing routine maintenance and capacity planning, the DBA today has to become a data expert with a key role in building foundations to transform organisations into high-performance businesses fit for the digital era. But to get there, organisations need to address how they use data, and DBAs must retool their query optimising skills to become data performance experts.
Rapid application development
Speed to market is essential for businesses to maintain a competitive advantage, and nowhere is this truer than in application development. Ultimately, the length of a development cycle can mean the difference between people hailing an app as the next Pokémon Go-style cultural phenomena or merely forgettable.
Under increased pressure to deliver applications at an ever-faster rate, developers now have the freedom to “do whatever it takes” to build at speed and ignore the constraints of traditional corporate IT processes. As such, the pressure to deliver faster has given rise to a whole new variety of database architectures such as document databases, which allow developers to compile data more quickly.
The document model works well when there are many variables that the developer wants to address. For example, consider a customer that takes their car to the garage for service. The customer record will contain information that is structured, such as billing information and the make and model of the car. Some of it may not be structured, because the mechanic may have taken ad hoc notes or pictures of the repair work to explain what has been done to fix it. This information could sit in a document database, enabling the developer to have all the information about the customer represented in a simple way, and information could be added to the database very quickly as the status of the car changes.
All this sounds great if you are an application developer, but it does make it harder to search and analyse data. Compare this to the old days when DBAs controlled the database structure to get the best out of it. The DBA was the data expert, who understood there are trade-offs and how to optimise systems to get the best performance overall. The DBA would make sure the database used the right indexes, normalised the schema and optimised the overall performance of queries. For instance, while spreading the information about a customer across several tables in a normalised schema may make the application developer’s job less straightforward when creating the record, a query to find everyone with an overdue bill could be highly optimised.
When not to optimise
Ultimately, knowledge concerning query optimisation lies with the DBA, as traditionally the role has always been about the data. It presents a great opportunity for DBAs to be integral to the performance of applications in the digital business era, but it does require an understanding of the whole toolbox of optimisation techniques available. The DBA’s role is broadening too, because it is not just about optimising queries as in days gone by; it is also about what that optimisation means for the business.
Application developers are going to continue to choose the database model that is most convenient for them, so it is incumbent on the DBA to optimise access to the data for users. That is more easily achieved if the DBA is using a multi-modal database that can integrate different data types, but also can be tuned for structured queries. It allows for far greater management and optimisation of the data in a single location.
The DBAs that become obsessive about such performance will be able to demonstrate the visible impact they are having on the business needs of their organisations. By its simplest measure, users will experience far more responsive, efficient, and accurate data searches, but more importantly, they will be able to do their jobs more effectively, collaborate more closely, and understand their customers more deeply as the data they are using is more complete. The DBAs that can meet this challenge will demonstrate how strategic their roles are to creating high-performance IT infrastructures for the digital business era.